Rapid Triage of Patients via Artificial Intelligence Processing of CT Images
Computerized tomography (CT) imaging in the Emergency Department (ED) critically affects patient diagnosis, triage, length of stay, and health outcomes. Currently, a radiologist must manually review all CT studies prior to patient diagnosis and triage, and several hours may pass between the time of the CT scan and the time of image evaluation and reporting by the radiologist. Image findings can be categorized as no acute (e.g. normal or chronic), acute (e.g. appendicitis), or hyperacute (e.g. rupture aortic aneurysm). We propose to develop an artificial intelligence (AI) algorithm that will rapidly analyze CT images and assist with triage of patients. What if the AI algorithm could identify hyperacute abnormalities that will kill in <1 hour and notify the radiologists and ED staff immediately? Conversely, what if we could identify the 30% of patients with no acute abnormality? We could prepare them for discharge and move them into the waiting room, freeing up beds/rooms for patients with life threatening emergencies. Did you know that 9 body regions make up 98.5% of all CT scans in the ED. We will attack each with great focus. This project is a collaboration between multiple UAB investigators and industry partners. I don't know of a more exciting project, and we have room for multiple students / trainees who can take on as little or as big of a component as they want. Want to launch your career and be a part of something great? Then you should join our team.
Contact Information
Andrew Smith MD PhD
andrewdennissmith@uabmc.edu
205-934-3138
